Medium Access Protocol Analysis -...
Transcript of Medium Access Protocol Analysis -...
-
Overview•Motivation•The performance of wireless multiple access systems is affected by different factors (e.g., channel, collisions, interference, etc.)•Current models emphasize a single factor •A wide variety Medium Access Control (MAC) protocols address only one factor
•Aim•Reverse engineer protocol performance in order to systematically forward engi-neer more efficient protocols for wireless multiple access systems
Error Event Approach
•Channel Model•We concentrate on a 2-user system trying to communicate to a single Base Station•There is a fading channel between each one of the users and the base station•Users access the medium following different protocols
•Analysis Procedure•In order to analyze the different MAC proto-cols, we propose the following procedure•1) Define what is an error•2) Identify distinguishable types of errors•3) Find what events cause errors (this de-pends on the system model)•4) Calculate the probability of the error-causing events•5) Analyze the effects of the protocolʼs reac-tion in the next step
•Graphical Representation•Our approach is suitable for graphical repre-sentation in the space of possible events at each system state•We introduce a diagram for each step in the protocol and it should convey the following information:•1) Probability of each event•2) Time required to reach that state•3) Amount of information transmitted if suc-cessful•4) Channel state•5) Power used in transmission
Analysis Examples
•Slotted ALOHA•Original model assumptions (coll. channel)
•Fading channel model
•802.11•Original model assumptions (coll. channel)
•Fading channel model
•Observations•The ALOHA protocol does not address any factor affecting the probability of error events•The 802.11 protocol reduces the possibility of back-to-back collisions, but was not designed to address the issue of fading channels•The approach can help identify unique situa-tions (e.g., collision with 2 ch. in deep fade)•Protocols can be improved by reducing the size of the error events as part of the protocol design
Conclusions•We have created a generalized analysis and design technique by evaluating proto-cols using error events•Our main contribution is the identification and aggregation of events into categories•From this viewpoint, we can make our models and analysis reflect real-life behaviors
2nd Annual North American School of Information Theory
Medium Access Protocol Analysis and Design: An Error-event Approach
Pedro Santacruz and Ashutosh Sabharwal(pesantacruz, ashu)@rice.edu
Base
Station
User
1
User
2
h1
h2
PEc PEc
e−λ(1− e−λ) e−λ(1− e−λ)
After successful Tx
PEc
After successful Tx
PEh1
PEh2
PEc
PEh1
PEh2
PEc = (1− e−λ)2
PEc = P′Ec − (P
′EcP
′Eh1 + P
′EcP
′Eh2 + 2P
′EcP
′Eh1P
′Eh2)
PEh1 = P′Eh1 − P
′EcP
′Eh1
PEh2 = P′Eh2 − P
′EcP
′Eh2
P ′Ec = (1− e−λ)2
P ′Eh1 = (1− e−λ)(1− exp
(−2R − 1SNR1
))
P ′Eh2 = (1− e−λ)(1− exp
(−2R − 1SNR2
))
PEcPEc
1−(
12W
)21−
(1W
)2
After successful Tx
PEcPEc
After successful Tx
PEh2
PEh1
PEh2
PEh1
t = 0 t = 1
t = 0 t = 1
P ′Ec(t) =(
12tW
)
P ′Eh2 = 1− exp(−2R − 1SNR2
)P ′Eh1 = 1− exp
(−2R − 1SNR1
)